REVISITING THE TOURISM-LED GROWTH HYPOTHESIS IN A DUAL MODEL USING MWALD GRANGER CAUSALITY ANALYSIS

Sung Yu-Chi¹

¹Department of Leisure & Tourism Management, Shu-Te University, Kaohsiung, Taiwan

Keywords:Co-integration Economic growth International tourist arrivals International tourist expenditures MWALD causality test Tourism-led growth hypothesis.

ARTICLE HISTORY: Received:28 March 2018. Revised:19 June 2018. Accepted:25 July 2018. Published:31 August 2018.

1. INTRODUCTION

Over the past decades, tourism has become one of the rapidly growing sectors of the world economy, and the contribution of tourism expansion to economic growth has drawn significant attention from a policy perspective. Taiwan has been ranked among the top four export-oriented economies in the Asian region (Jin and Shih, 1995 ) and had maintained its status of being among the top two until the year of 2000. In 2002, the Taiwanese government recognized the importance of the tourism sector in the economy, and came to believe that tourism expansion may gradually lead to potential high economic growth. Thus, the Taiwanese government considered it worthwhile to introduce some useful polices to stimulate tourism activities. For example, the Taiwanese government attempted to bring about a significant change in tourism policy by relaxing tourist regulations since Taiwan’s separation from Mainland China in 1949. Thereafter, the Congress introduced the Doubling Tourist Arrivals Plan (DTAP) to release a series of national development plans entitled “Challenge 2008,” which was designed to promote the tourism industry. According to the Annual Statistics of Tourism 2012, Taiwan’s international tourist arrivals accounted for 5.3 percent of the Taiwan’s GDP in 2004, which increased to 7.8 percent in 2009. Despite major global events such as the terrorist attacks that took place in the US on September 11, 2001 and the global financial crisis in 2008 and 2009, which led to a temporary global economic recession in the following two years, tourism expansion in Taiwan has been stable. This may unveil the tourism industry’s crucial contribution to Taiwan’s economic growth. In addition, according to the latest United Nations World Tourism Organization (UNWTO) report, international tourist arrivals have consistently grown by 4% and have remarkably grown by 7% since 2017. Furthermore, international tourist expenditure has been increasing steadily since 2010. This increasing tendency is expected to continue in 2018 at a rate of approximately 5% to 6%.

It is commonly believed that tourism expansion has a long-run positive effect on economic growth, and that it promotes economic growth in several ways. First, tourism development is necessary for the accumulation of foreign exchange earnings for developing countries, which in turn can be used to finance the imported capital goods utilized in the production process (Tomohara, 2016 ). Second, tourism activities can benefit other sectors of the economy such as construction, transportation, and hospitality, as well as cities of environmental improvement. Third, a successful tourism industry has the ability to create market demand and job opportunities, which in turn leads to higher household income and tax revenue for the government through augmented multiplier effects (Tang et al., 2015 ; Liu et al., 2017 ). Finally, the tourism industry can be considered as an imperative factor in the transmission of knowledge economy through human capital development. In other words, higher economic growth can be achieved not only by improving the factors of labor and capital but also by escalating tourism expansion. Therefore, the mutual relationship between tourism expansion and economic growth must be considered by policymakers (Balaguer and Cantavella-Jorda, 2002 ; Matarrita-Cascante, 2010 ) and the tourism-led growth hypothesis is worthy of analysis.

A growing number of studies in the tourism literature have discussed the relationship between tourism expansion and economic growth, using the causality test developed by Granger (1969 ). The existing literature on this topic has put forward four hypotheses. The first hypothesis is the so-called “tourism-led economic growth” hypothesis that assumes tourism boosts economic growth, which results from the economic benefits of tourism. This hypothesis interprets that a unidirectional causality runs from tourism expansion to economic growth, which has been supported by;  Balaguer and Cantavella-Jorda (2002 ); Lanza et al. (2003 ); Carrera et al. (2008 ); Ghartey (2013 ) and Tang and Tan (2015 ); Mishra and Rout (2016 ) and Ohlan (2017 ). The second hypothesis is the “economic-driven tourism growth” hypothesis (Narayan, 2004 ; Oh, 2005 ; Payne and Mervar, 2010 ) which asserts that economic growth may contribute to tourism expansion, and that a unidirectional causality runs from economic growth to tourism expansion. The third hypothesis is the so-called “reciprocal or feedback” hypothesis (Durbarry, 2004 ; Kim et al., 2006 ; Lee and Chang, 2008 ; Corrie, 2013 ; Shahzad et al., 2017 ) which postulates a bidirectional causality between economic growth and tourism expansion . This hypothesis observes a reciprocal relationship between economic growth and tourism expansion. Specifically, increasing the economic growth of a country fosters the development of tourism-related sectors in the domestic country as potential resources, which is available for any kind of tourism infrastructure, and any proliferation of tourism activities may lead to international exchange income accumulation, which feeds back to economic performance. The fourth hypothesis is the “no-causality” hypothesis, which demonstrates a neutral relationship between economic growth and tourism expansion. This hypothesis has been supported by Eugenio-Martın and Morales (2004 ); Chen and Chiou-Wei (2009 ); Katircioglu (2009 ) and Brida et al. (2010 ) for different countries.

The main purpose of this study is to reexamine the tourism-led growth hypothesis by incorporating the factor of international tourist expenditures into the MWALD causality test (Toda and Yamamoto, 1995 ) to enhance the integrity of empirical results. So far, most of the existing literatures usually consider the case of multiple countries to take advantage of numerous data, in order to facilitate comparison of the causality analysis; however, their empirical results are undetermined owing to missing data for some countries or inappropriate model selection. In this study, in an attempt to improve the empirical credibility of our findings, we construct two models (bivariate and trivariate models) to compare the long-run and causality relationships between variables. We also use sufficient and continuous year data for a single country, and consider both the international tourists arrivals and expenditures simultaneously to analyze the tourism-economic growth nexus.

The remainder of this study is organized as follows. Section 2 discusses the data and introduces the econometric methods, including the Johansen co-integration test and MWALD causality tests. Section 3 reports the empirical results. Finally, Section 4 concludes the study and discusses policy implications.

2. METHODOLOGY AND DATA

2.1. Model Classification

To compare the difference between the tourism-led growth and the existence of co-integration, we employ the bivariate and trivariate models, which are denoted as Eq. (1) and Eq. (2), respectively.

 All the three variables are measured in US million dollars and taken as logarithms before conducting the empirical analysis. We used long-term time series data ranging from 1958 to 2017. The data sources are the World Bank 2017, the Directorate General of Budget, Accounting and Statistics, Executive Yuan, Taiwan, and the tourism statistics database of the Taiwan Tourism Bureau.

2.2. Unit Root Test - Augmented Dickey Fuller (ADF)

To avoid the spurious problem of regressions, we use the Augmented Dickey–Fuller (ADF) test to examine the stationarity among the selected variables (Granger and Newbold, 1974 ). The ADF test detects the explanatory variables’ serial correlation of their lagged values from the following regression:

2.3. Johansen’s Co-Integration Test

Johansen and Juselius (1990 ) show that a long-run co-integrating relationship may exist among variables once the stationarity is confirmed in data sets. In this study, we apply the Johansen multivariate co-integration test to

The reason for applying the Johansen co-integration test is that when the integration orders of variables are all in consistency, the Johansen co-integration test can be performed accurately, even though the error term is non-normal distribution or the lag terms in the vector error-correction model (VECM) are mislaid (Gonzalo, 1994 ; Clarke and Mirza, 2006 ).

To investigate the long-run co-integration relationship, we employ the VECM model along with Eq. (3) as follows:

2.4. Modified Wald (MWALD) Causality Test

While the Johansen multivariate co-integration approach uncovers the long-run co-integration information, it does not reveal the causal relationships. In this study, we construct an augmented Vector Autoregression (VAR) model along with the MWALD causality test developed by Toda and Yamamoto (1995 ) in order to explore the tourism-led growth hypothesis. The reasons for using the MWALD causality test are that it is simple to use and the pre-testing for the co-integration properties is not required as long as the estimation of the augmented VAR assures the asymptotic distribution of the Wald test. Therefore, the MWALD causality test is adaptive regardless of whether the arbitrary integrated orders of variables, non-co-integrated or co-integrated (Zapata and Rambaldi, 1997 ). To implement the MWALD causality test, we incorporate the augmented VAR models into the bivariate model of Eq. (7) and trivariate model of Eq. (8) as follows:

3. EMPIRICAL RESULTS

3.1. Results of Unit Root Test

Thus, the null hypothesis of unit root is rejected at the 10% significance level, which suggests that all the three variables are in their stationarity when taking the first difference. In other words, these variables are integrated of order one, I(1), which are consistent with the findings of Nelson and Plosser (1982 ). Therefore, the confirmation of stationarity helps us continue the Johansen co-integration test to examine if there exists a long-run relationship among variables.

Table-1. ADF unit root test results

Note: Numbers in parentheses refer to the selected lag orders determined by the AIC rule. ** and *** represents the null is rejected under the 5% and 1% level, respectively.

3.2. Results of Johansen Co-Integration Test

Table-2. Johansen co-integration test results

Note: Numbers in parentheses of models refer to the optimal lag orders determined by the AIC rule. r is the co-integration rank. ** and *** represents the null is rejected under the 5% and 1% significant level, respectively, and the critical values are obtained from the asymptotic critical values (Osterwald‐Lenum, 1992 ).

3.3. Results of MWALD Causality Test

According to the Granger causality theory (Granger, 1969 ) if the variables are in their stationarity and co-integrated, there should be at least one unidirectional causality relationship among them.

Table 3 shows the MWALD causality test results for the bivariate model and indicates that at the 5% significance level, the null hypothesis that the real GDP does not Granger-cause international tourist arrivals is rejected. This suggests that the tourism-led growth hypothesis is invalid and there is only a unidirectional Granger causality running from real GDP to international tourist arrivals in the bivariate model.

Table-3. MWALD causality test for the bivariate model

Causality direction	Lag order	MWALD statistics	P-Value
lnTA does not cause lnY	9	7.18	0.24
lnY does not cause lnTA	8	18.31**	0.05

Note: ** and *** indicate significance at the 5% and 1% level, respectively, and the optimal lag order is determined by the AIC rule.

In Table 4, the trivariate model presents two different results. First, the null hypothesis that the real GDP does not Granger-cause international tourist arrivals is rejected at the 1% significance level. This implies that there is a unidirectional Granger causality running from real GDP to international tourist arrivals. Second, the null hypothesis that the international tourist expenditures (GDP) does not Granger-cause real GDP (international tourist expenditures) is rejected at the 5% (1%) significance level. This result shows that there is a bidirectional Granger causality between real GDP and international tourist expenditures. Compared to the bivariate model, the tourism-led growth hypothesis is strictly approved in the trivariate model based on the above causal results analysis.

Table-4. MWALD causality test for the trivariate model

Causality direction	Lag order	MWALD statistics	P-Value
lnTA does not cause lnY	7	9.03	0.25
lnY does not cause lnTA	9	28.64***	0.01
lnTE does not cause lnY	6	17.01**	0.05
lnY does not cause lnTE	8	46.02***	0.00

Note: ** and *** indicate significance at the 5% and 1% level, respectively, and the optimal lag order is determined by the AIC rule.

4. CONCLUSION AND POLICY IMPLICATIONS

This study employed the Johansen co-integration and MWALD causality tests to examine the relationships among the real GDP, international tourist arrivals, and international tourist expenditures in Taiwan. Additionally, we re-investigated the tourism-led growth hypothesis by analyzing both the bivariate and trivariate models, using empirical techniques. First, we applied the ADF test to ensure the stationarity and integration orders of all variables. Thereafter, we employed the Johansen co-integration approach to discover the long-run relationship between the international tourist arrivals (international tourist expenditures) and the real GDP. Second, we constructed an augmented VAR model in both the bivariate and trivariate models to examine the MWALD causal relationship among these variables.

The main findings of the empirical analysis were as follows: (1) The unit root test results verified the stationarity of all variables; (2) The Johansen co-integration test results provided evidence that all variables are co-integrated and there is a long-run relationship among variables in both bivariate and trivariate models; (3) The MWALD causality test results show that there is a unidirectional Granger causality running from real GDP to international tourist arrivals in both bivariate and trivariate models, supporting the tourism-led growth hypothesis. In addition, there is bidirectional Granger causality between real GDP and international tourist expenditures only in the trivariate model.

On the basis of the empirical results, we can put forth two perspectives that may offer policymakers a better understanding of tourism- growth nexus to formulate tourism enticements in Taiwan. First, from the perspective of the tourism expansion policy, the bidirectional Granger causality has already provided evidence that sufficient international tourist expenditures may lead to accumulation of international reserves through exchange rate and further enhance economic growth. On the contrary, higher economic growth implies that more resources can be allocated to tourism through government spending. Therefore, policymakers could consider how to uncover the causal relationship between tourism and economic growth by formulating appropriate tourism policies. For example, by helping the tourism industry expand tremendously, and meanwhile, strengthening the country’s economic vitality such as public infrastructure, financial aid of loanable funds, and services training to attract international tourists’ travelling and incentives. Second, from the perspective of economic growth, by scrutinizing the tourism expansion policy, the feasible way to develop tourism is to pay more attention to attracting international tourists. For example, negotiate issues for abolishing the traveling visa limitation; shorten visa application procedure, increase the variety of free-tax commodity, or specify traveling expenditure discount for international tourists. In addition, the Granger causality evidence in both bivariate and trivariate models suggests that promoting economic growth as its top priority is an useful policy and this also indirectly supports the tourism-led growth hypothesis.

Funding: This study received no specific financial support.

Competing Interests: The author declares that there are no conflicts of interests regarding the publication of this paper.

 REFERENCES

Balaguer, J. and M. Cantavella-Jorda, 2002. Tourism as a long-run economic growth factor: The Spanish case. Applied economics, 34(7): 877-884. View at Google Scholar | View at Publisher

Brida, J.G., B. Lanzilotta, S. Lionetti and W.A. Risso, 2010. Research note: The tourism-led growth hypothesis for Uruguay. Tourism Economics, 16(3): 765-771.View at Google Scholar | View at Publisher

Carrera, E.J.S., W.A. Risso and J.G. Brida, 2008. Tourism's impact on long-run Mexican economic growth. Economics Bulletin, 23(21): 1-8. View at Google Scholar | View at Publisher

Chen, C.-F. and S.Z. Chiou-Wei, 2009. Tourism expansion, tourism uncertainty and economic growth: New evidence from Taiwan and Korea. Tourism Management, 30(6): 812-818.View at Google Scholar | View at Publisher

Clarke, J.A. and S. Mirza, 2006. A comparison of some common methods for detecting Granger noncausality. Journal of Statistical Computation and Simulation, 76(3): 207-231. View at Google Scholar | View at Publisher

Corrie, K., 2013. Tourism and economic growth in Australia: An empirical investigation of causal links. Tourism Economics, 19(6): 1317-1344. View at Google Scholar | View at Publisher

Dolado, J.J. and H. Lütkepohl, 1996. Making wald tests work for cointegrated var systems. Econometric Reviews, 15(4): 369-386.View at Google Scholar | View at Publisher

Durbarry, R., 2004. Tourism and economic growth: The case of Mauritius. Tourism Economics, 10(4): 389-401. View at Google Scholar | View at Publisher

Eugenio-Martın, J.L. and N.M. Morales, 2004. Tourism and economic growth in Latin American countries: A panel data approach. Fondazione Eni Enrico Mattei (FEEM), Working Paper, No. 26.

Ghartey, E.E., 2013. Effects of tourism, economic growth, real exchange rate, structural changes and hurricanes in Jamaica. Tourism Economics, 19(4): 919-942.View at Google Scholar | View at Publisher

Gonzalo, J., 1994. Five alternative methods of estimating long run equilibrium relationship. Journal of Economics, 60(1): 202-233. View at Google Scholar | View at Publisher

Granger, C.W. and P. Newbold, 1974. Spurious regressions in econometrics. Journal of Econometrics, 2(2): 111-120. View at Google Scholar | View at Publisher

Granger, C.W.J., 1969. Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3): 424-438. View at Google Scholar | View at Publisher

Jin, J.C. and Y. Shih, 1995. Export-led growth and the four little dragons. Journal of International Trade & Economic Development, 4(2): 203-215. View at Google Scholar | View at Publisher

Johansen, S., 1991. Estimation and hypothesis testing of co-integrating vector in Gaussian vector autoregression models. Econometrica, 59(6): 1551-1580. View at Google Scholar | View at Publisher Johansen, S., 1994. The role of the constant and linear terms in co-integration analysis of stationary variables. Econometric Reviews, 13(2): 205-229. View at Google Scholar | View at Publisher

Johansen, S. and K. Juselius, 1990. Maximum likelihood estimation and inference on cointegration—with applications to the demand for money. Oxford Bulletin of Economics and Statistics, 52(2): 169-210. View at Google Scholar | View at Publisher

Katircioglu, S.T., 2009. Revisiting the tourism-led-growth hypothesis for Turkey using the bounds test and Johansen approach for cointegration. Tourism Management, 30(1): 17-20. View at Google Scholar | View at Publisher

Kim, H.J., M.H. Chen and S.C. Jang, 2006. Tourism expansion and economic development: The case of Taiwan. Tourism Management, 27(5): 925-933.View at Google Scholar | View at Publisher

Lanza, A., P. Temple and G. Urga, 2003. The implications of tourism specialisation in the long run: An econometric analysis for 13 OECD economies. Tourism Management, 24(3): 315-321.View at Google Scholar | View at Publisher

Lee, C.C. and C.P. Chang, 2008. Tourism development and economic growth: A closer look at panels. Tourism Management, 29(1): 180 -192. View at Google Scholar | View at Publisher

Liu, J., P. Nijkamp and D. Lin, 2017. Urban-rural imbalance and tourism-led growth in China. Annals of Tourism Research, 64: 24-36.View at Google Scholar | View at Publisher

Matarrita-Cascante, D., 2010. Beyond growth: Reaching tourism-led development. Annals of Tourism Research, 37(4): 1141-1163. View at Google Scholar 

Mishra, P.K. and H.B. Rout, 2016. Tourism in Odisha: An engine of long run growth. Journal of Tourism Management Research, 3(2): 74-84.View at Google Scholar | View at Publisher

Narayan, P.K., 2004. Economic impact of tourism on Fiji's economy: Empirical evidence from the computable general equilibrium model. Tourism Economics, 10(4): 419-433. View at Google Scholar | View at Publisher

Nelson, C.R. and C.R. Plosser, 1982. Trends and random walks in macroeconmic time series: Some evidence and implications. Journal of Monetary Economics, 10(2): 139-162. View at Google Scholar | View at Publisher

Oh, C.-O., 2005. The contribution of tourism development to economic growth in the Korean economy. Tourism Management, 26(1): 39-44. View at Google Scholar | View at Publisher

Ohlan, R., 2017. The relationship between tourism, financial development and economic growth in India. Future Business Journal, 3(1): 9-22. View at Google Scholar | View at Publisher

Osterwald‐Lenum, M., 1992. A note with quantiles of the asymptotic distribution of the maximum likelihood cointegration rank test statistics. Oxford Bulletin of Economics and Statistics, 54(3): 461-472. View at Google Scholar | View at Publisher

Payne, J.E. and A. Mervar, 2010. Research note: The tourism–growth nexus in Croatia. Tourism Economics, 16(4): 1089-1094.View at Google Scholar | View at Publisher

Shahzad, S.J.H., M. Shahbaz, R. Ferrer and R.R. Kumar, 2017. Tourism-led growth hypothesis in the top ten tourist destinations: New evidence using the quantile-on-quantile approach. Tourism Management, 60: 223-232.View at Google Scholar | View at Publisher

Tang, C.F., Y.W. Lai and I. Ozturk, 2015. How stable is the export-led growth hypothesis? Evidence from Asia's four little dragons. Economic Modelling, 44: 229-235.View at Google Scholar | View at Publisher

Tang, C.F. and E.C. Tan, 2015. Tourism-led growth hypothesis in Malaysia: Evidence based upon regime shift cointegration and time-varying Granger causality techniques. Asia Pacific Journal of Tourism Research, 20(sup1): 1430-1450. View at Google Scholar | View at Publisher

Toda, H.Y. and T. Yamamoto, 1995. Statistical inference in vector autoregressions with possibly integrated processes. Journal of Econometrics, 66(1-2): 225-250.View at Google Scholar | View at Publisher

Tomohara, A., 2016. Japan's tourism-led foreign direct investment inflows: An empirical study. Economic Modelling, 52: 435-441. View at Google Scholar 

Zapata, H.O. and A.N. Rambaldi, 1997. Monte Carlo evidence on cointegration and causation. Oxford Bulletin of Economics and Statistics, 59(2): 285-298.View at Google Scholar | View at Publisher